Global environmental change is challenging our understanding of how communities as a whole interact with their physical environment. Ideally, we would model the impacts of global environmental change at a global level. However, in order to mathematically model the sheer functional diversity of Earth's dynamic ecosystems, we need to integrate the scales at which these processes operate. Traditionally, studies of ecosystem function have focused on singular ecological, evolutionary or biogeochemical process within an environment. Such studies have contributed much more to the development of our understanding of ecosystem function than those focused on the interactions between biotic and abiotic factors. Ultimately, the productivity of most ecosystems is controlled by the concentration, molecular form and stoichiometry of the macronutrients thereby highlighting the importance of biogeochemical modelling for dynamic ecosystem models across molecular, habitat, landscape and global scales. But as we face unprecedented rates of habitat degradation and species extinctions, few traditional theories can predict in detail how ecosystems will respond to perturbations such as environmental disturbance or shifting weather patterns. To be both statistically and ecologically informative, future ecosystem and biogeochemical models must address complex interactions from atoms to ecosystems. Unless ecological processes are modelled explicitly, significant feedbacks, thresholds and constraints will be missed. The aim of this chapter is to review the state of the art in the use of such models, and suggest new approaches for ecologists, biogeochemists and mathematicians to work together to model the inputs and outputs of entire ecosystems rather than as a series of individual interactions.